Ballasting effects of smectite on aggregate formation and export from a natural plankton community

Marine Chemistry 175 (2015) 18–27

Contents lists available at ScienceDirect

Marine Chemistry journal homepage: www.elsevier.com/locate/marchem

Ballasting effects of smectite on aggregate formation and export from a natural plankton community Morten H. Iversen a,b,c,⁎, Maya L. Robert a,1 a b c

Alfred Wegener Institute for Polar and Marine Research, Am Handelshafen 12, 27570 Bremerhaven, Germany Helmholtz Young Investigator Group SEAPUMP, Alfred Wegener Institute for Polar and Marine Research, Am Handelshafen 12, 27570 Bremerhaven, Germany Faculty of Geosciences and MARUM, University of Bremen, Klagenfurter and Leobener Strasse, 28359 Bremen, Germany

a r t i c l e

i n f o

Article history: Received 31 August 2014 Received in revised form 27 April 2015 Accepted 29 April 2015 Available online 1 May 2015 Keywords: Smectite Ballast minerals Ballast hypothesis Particulate organic carbon export Size-specific sinking velocity Carbon-specific respiration rate Aggregates

a b s t r a c t Strong correlations between particulate organic carbon (POC) and ballast minerals have been observed in the deep ocean. This has led to the postulation that ballast minerals can enhance POC flux by increasing the density and sinking velocity of ballasted aggregates and/or that ballast minerals protect the aggregated organic matter from degradation. Here we experimentally tested the influence of the ballast mineral smectite on the formation, size, dry weight, size-specific sinking velocity, carbon-specific respiration rate, and total POC flux of marine snow aggregates formed in roller tanks from a natural plankton community isolated from the North Sea. This study shows that the inclusion of smectite offers no protection against degradation of organic matter in freshly produced or aged marine snow aggregates. The main effect of ballasting with smectite was an increase in the density of the aggregates and, therefore 2- to 3-fold higher size-specific sinking velocities. Mineral ballasting had no influence on the total volume of aggregates or the total aggregated amount of POC. Nevertheless, the effect of increased sinking velocities in the ballasted treatment resulted in 2.7 ± 1.6 times larger potential POC fluxes compared to the non-ballasted aggregates. This implies that the incorporation of ballast minerals into sinking organic aggregates can increase the efficiency of the biological pump. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The biological carbon pump refers to the biologically driven processes that transport particulate organic matter from the surface ocean to the deep sea and sediments (Falkowski et al., 1998). Most of the exported particulate organic matter is transported in the form of organic settling particles such as fecal pellets and macroscopic aggregates (N500 μm diameter) known as marine snow (Fowler and Knauer, 1986). The efficiency of the biological pump is controlled by several factors, including aggregation (e.g. Alldredge and Gotschalk, 1988; Alldredge et al., 1995; Fowler and Knauer, 1986) and disaggregation (e.g. Alldredge et al., 1990) of the settling aggregates, their sinking velocities (e.g. Alldredge and Gotschalk, 1988), microbial remineralization (e.g. Grossart and Ploug, 2001; Iversen and Ploug, 2010; Ploug, 2001), as well as zooplankton consumption (e.g. Iversen et al., 2010; Jackson, 1993; Koski et al., 2005) and transformation (e.g. Dilling and Alldredge, 2000; Iversen and Poulsen, 2007; Lampitt et al., 1990). Additionally, deep ocean sediment

⁎ Corresponding author at: Alfred Wegener Institute for Polar and Marine Research, Am Handelshafen 12, 27570 Bremerhaven, Germany. E-mail address: [email protected] (M.H. Iversen). 1 Present address: 285 Chemin Peymeyan, 83440, St Paul en Forêt, France.

http://dx.doi.org/10.1016/j.marchem.2015.04.009 0304-4203/© 2015 Elsevier B.V. All rights reserved.

trap studies (Deuser et al., 1981) show a strong correlation between fluxes of particulate organic carbon (POC) and minerals such as biogenic silica, calcium carbonate and lithogenic minerals. This has led to the “ballast hypothesis”, which suggests that an association between minerals and organic matter within sinking particles controls the export of POC to the deep ocean (Armstrong et al., 2002; Francois et al., 2002; Klaas and Archer, 2002). It has been suggested that the minerals increase the densities of the sinking aggregates and, thus, increase their sinking velocities and/or that the minerals provide protection of the organic matter against remineralization (Armstrong et al., 2002; Engel et al., 2009a; Fischer and Karakas, 2009; Francois et al., 2002; Hedges et al., 2001; Iversen and Ploug, 2010; Klaas and Archer, 2002; Lee et al., 2009a; Passow et al., 2003; Ploug et al., 2008a; Sanders et al., 2010). The ballast hypothesis has spawned a series of publications investigating the role of ballast minerals for the export of POC. Investigations on the protection of settling aggregates against remineralization via the incorporation of ballast minerals have found evidence supporting both outcomes, not only i) the presence of a protective mechanism (Arnarson and Keil, 2000; Engel et al., 2009a; Le Moigne et al., 2013) but also ii) the lack of an effect on remineralization (Ingalls et al., 2006; Iversen and Ploug, 2010; Ploug et al., 2008a,b). Several studies have confirmed the potential for ballast minerals to increase the density and, thus, the sinking velocity of aggregates (e.g. De La Rocha et al.,

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27

2008; De La Rocha and Passow, 2007; Engel et al., 2009a; Iversen et al., 2010; Iversen and Ploug, 2010; Ploug et al., 2008a,b; Thomalla et al., 2008). Recent studies have observed that the presence of ballast minerals can enhance POC fluxes, confirming the ballasting effect of minerals (Bressac et al., 2014; Lee et al., 2009b; Sanders et al., 2010; Ternon et al., 2010; Thunell et al., 2007). This shows that the ballasting effect could have important consequences for the future climate, in which it has been suggested that increased desertification and droughts will occur, which could lead to higher dust availability in the atmosphere and ocean (e.g. Mahowald et al., 2009; Prospero and Nees, 1976). Despite the seemingly clear effect of ballasting toward denser aggregates with larger size-specific sinking velocities, only few studies have been undertaken on the overall effect of ballasting on total export fluxes (Bressac et al., 2014; Ternon et al., 2010). Both studies observed an enhanced export flux in the presence of ballast minerals, but did not combine their results with direct measurements on individual aggregates, i.e. the influence of ballasting on either sinking velocity or degradation of individual aggregates. There is therefore a need to improve our understanding of how ballast minerals can induce enhanced export fluxes on both the scale of individual aggregates and the overall effect on the total aggregated material: i.e. i) will the increase in aggregate abundance typically observed for ballasted aggregates compensate for the mass-loss due to decreased maximum aggregate size (Hamm, 2002; Iversen and Ploug, 2010; Passow and De La Rocha, 2006), and ii) will the increase in excess density cause the smaller aggregates to sink as fast or faster than the large non-ballasted aggregates and, thus, enhance the total POC flux in comparison to non-ballasted aggregates? This study aims to improve our understanding of the causality of ballast minerals in enhancing POC fluxes by concomitantly measuring ballast effects on individual aggregate parameters such as size, sinking velocity, dry weight, solid hydrated density, particulate organic carbon and nitrogen content, and microbial respiration rate of ballasted and non-ballasted aggregates over time. Additionally, we will follow the total pool of aggregates in terms of abundance of different sizes, total aggregated volume, and total potential POC flux in both a ballasted and a non-ballasted treatment. We chose smectite as ballasting mineral since this is often an important component of desert dust, e.g. Saharan dust (Avila et al., 1996), which has been shown to enhance POC flux in both the Mediterranean (Bressac et al., 2014; Ternon et al., 2010) and in the second largest Eastern Boundary upwelling area off Cape Blanc, Mauritania (Iversen et al., 2010). Additionally, smectite, together with illite, is one of the most abundant minerals in situ (Weaver, 1988) and is the predominant clay mineral in the world's oceans (Cole and Shaw, 1983). 2. Material and methods 2.1. Aggregate formation 100 L of North Sea water (salinity 31.6 and temperature 15 °C) was collected from the coast off Helgoland at the Helgoland Roads Time Series site (54.11°N, 7.54°E) on 29 April 2007 and carefully inversefiltered through a 100 μm mesh to remove mesoplankton. Sampling coincided with the measured fluorescence maximum at the spring peak, with chlorophyll a concentrations of ~5 μg L−1. The in situ chlorophyll a concentrations at the sampling site decreased to ~2 μg L−1 only a few days after the sampling, suggesting that the plankton community was at a late bloom stage (see Löder et al., 2011 for more information). On 30 April 2007, the b100 μm plankton community was incubated in 35 Plexiglas cylinders, each with a volume of 1.15 L (roller tanks, 14 cm diameter and 7.47 cm height). Qualitative microscopic observations of the b100 μm phytoplankton community showed that it was dominated by diatoms: the most abundant species was Chaetoceros sp., while Skeletonema sp., Thalassiosira sp., and Coscinodiscus sp. were

19

also present at high abundances. Smectite clay (Montmorillonite: SWy-2, Clay Mineral Society, Colorado, USA) was added with a final concentration of 0.5 mg L−1 to 15 of the roller tanks to act as a ballast mineral while the remaining 20 roller tanks were non-ballasted. To our knowledge no measurements of surface ocean concentrations of smectite are known. We selected this concentration to imitate mineral input at concentrations similar to previous ballast mineral studies in roller tanks with clay additions in the range of 5 to 100 mg L− 1 (Hamm, 2002) or 0.007 to 10 mg L− 1 (Passow and De La Rocha, 2006). A mesocosm study in the Mediterranean Sea added 41.5 g of Saharan dust to 52 m3 sea water, i.e. 0.8 mg of Saharan dust per L− 1 (e.g. Bressac et al., 2011, 2014; Guieu et al., 2014). Since 25% of the Saharan dust was composed of clay particles, their clay addition was 0.2 mg L−1, which is similar to the one used in this study. Scanning electron microscope images of the non-ballasted incubations did not show any presence of smectite. All 35 roller tanks were placed on a roller table and rotated with 3 RPM at 15 °C in darkness. The aggregate dynamics were followed throughout the study by counting the number of aggregates within different size classes (b1 mm, 1–2 mm, 2–3 mm, 3–4 mm, 4–5 mm and N5 mm) in each roller tank. 2.2. Sinking velocity Measurements were carried out at 5 sampling time points: 48, 91, 157, 187 and 380 h after incubation start. On each sampling day, 15 to 25 individual aggregates from each treatment (ballasted and nonballasted) were gently transferred with a wide bore pipette from the roller tanks to a vertical flow system where the sinking velocity of each aggregate was measured (see Ploug and Jørgensen, 1999). Generally, the aggregates from two randomly selected roller tanks were used for measurements at each sampling time point. This decreased the total number of roller tanks in each treatment by two after each sampling time point. The water in the vertical flow system was GF/F filtered and had the same temperature and salinity as the water in the roller tanks. An upward flow was adjusted to balance the aggregate sinking velocity until the aggregate remained suspended at a distance of one aggregate diameter above the net placed in the middle of the flow chamber (see Ploug et al., 2010). The sinking velocity of an aggregate was calculated from the flow rate divided by the cross-sectional area of the flow chamber. The x-, y-, and z-axes of each aggregate were measured in the flow system using a horizontal dissection microscope with a calibrated ocular. The aggregate volume was calculated by assuming an ellipsoid shape. For comparison with other aggregate shapes we calculated the equivalent spherical diameter (ESD) of each aggregate. 2.3. Oxygen measurements Oxygen concentrations were measured in 50 μm increments across the aggregate–water interface using a Clark-type oxygen microelectrode with a guard cathode (Revsbech, 1989). The 90% response time of the electrode was b 1 s and the stirring sensitivity b0.3%. The oxygen microelectrode was mounted in a micromanipulator and calibrated at air-saturation and at anoxic conditions. The electrode current was measured on a picoamperemeter (Unisense, PA2000) and read on a strip chart recorder (Kipp and Zonen) at high resolution (2 μM O2 cm− 1). The tip diameter of the microsensor was 2 μm. All measurements were done at the steady state of the oxygen gradients. See Iversen and Ploug (2010, 2013) for further details. 2.4. Total microbial respiration rates within the aggregates Oxygen fluxes to the aggregate and total respiration rates of the microbial community within the aggregates were calculated from the oxygen gradients measured across the aggregate–water interface under steady-state conditions (see Ploug et al., 1997). We used

20

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27

temperature- and salinity-corrected oxygen diffusion coefficients of 1.72 × 10− 5 cm2 s− 1 in the calculations (Broecker and Peng, 1974). The surface area of an ellipsoid (Maas, 1994) was used to calculate total oxygen consumption. The oxygen consumption rates were converted to carbon respiration by assuming a respiratory quotient of 1 mol of O2 to 1 mol of CO2, as also used in previous studies of oxygen respiration and POC degradation in diatom aggregates (e.g. Iversen and Ploug, 2013; Ploug and Grossart, 2000). 2.5. Aggregate dry weight and particulate organic carbon and nitrogen content On each sampling day, 5 to 10 individual aggregates of known size were filtered onto pre-weighed 0.2 μm polycarbonate filters to determine the aggregate dry weight (DW). Each filter contained one aggregate and was washed with de-ionized water to remove salt before drying at 40 °C for 48 h. The dry filters were weighed on a Mettler Toledo (UMX 2) balance with a sensitivity of 0.1 μg. The ratio of particulate organic carbon (POC) to DW was determined by filtering 50 aggregates onto preweighed 25 mm GF/F filters. The filters were gently rinsed with deionized water, and dried at 40 °C for 48 h before being reweighed on a Mettler Toledo (UMX2) balance. The GF/F filters were then fumed with hydrochloric acid and the POC and particulate organic nitrogen (PON) contents of the aggregates on each filter measured on an EA mass spectrometer (ANCA-SL 20-20, Sercon Ltd. Crewe, UK) with a precision of ±0.7 μgC or 0.3%. On each sampling day we determined the POC:DW ratio and the POC:PON ratio for both treatments. The measured POC:DW of the aggregates measured at the different sampling days were used to estimate POC content within the measured aggregates by multiplying the DW of each aggregate by the POC:DW ratio.

calibrated according to standard beads of different sizes that were analyzed on the flow cytometer between each sample run. 2.7. Excess density Navier–Stokes drag equation was used to calculate the excess density (Δρ) of the aggregates in each treatment for each sampling day (Stokes, 1851):

Δρ ¼

C D ρw w 2 ; 4 gESD 3

ð1Þ

where CD is the dimensionless drag force defined in Eq. (2), ρw is the density of seawater (1.0237 g cm−3, for a salinity of 32 at 15 °C), w is the measured sinking velocity in cm s−1, g is the gravitational acceleration of 981 cm s−2, and ESD is the equivalent spherical diameter in cm. We calculated CD using the drag equation for Re N1 given by White (1974):

CD ¼

    24 6 þ þ 0:4; 0:5 Re 1 þ Re

ð2Þ

where the Reynolds number (Re) was defined as Re ¼ w ESD

ρw ; η

ð3Þ

where η is the dynamic viscosity (1.2158 × 10−2 g cm−1 s−1 for a salinity of 32 at 15 °C).

2.6. Free-living and aggregate-associated microbial abundances 2.8. Solid hydrated density of aggregate constitutes The abundances of aggregate-associated and free-living bacteria, phytoplankton and flagellates were measured over time with a flow cytometer (BD FACSCalibur™). The aggregate-associated bacteria, flagellates and phytoplankton were detached by following the method developed by Lunau et al. (2005). The samples for the free-living microorganisms were taken from roller tanks after allowing the aggregated material to settle out for a minimum of 30 min. At each time point, individual aggregates were transferred from the roller tank to the flow chamber for size determination and sinking velocity. Since the flow chamber was filled with GF/F filtered seawater, this procedure removed any free-living organisms transferred together with the aggregate to the flow chamber. After determination of size and sinking velocity, each aggregate was placed in a 2 ml Eppendorf capped microcentrifuge tube (Eppendorf, Hamburg, Germany) to which 1500 μl sterile filtered seawater with formaldehyde (final concentration 4%) was added. Additionally, 150 μl of 100% methanol was added to the Eppendorf tube (10% (v/v) final concentration) to break up the exopolymeric substances and detach the aggregate-associated microorganisms (Lunau et al., 2005). The Eppendorf tubes were then incubated for 15 min at 35 °C in an ultrasonic bath (35 kHz). The methanol treatment was not applied to measurements of the free-living microbes. Each sample was centrifuged at 5 °C and 1250 relative centrifugal force (2000 rotations per minute) for 30 s to remove smectite particles from the supernatant. This procedure was also followed for the nonballasted treatment to ensure comparison and to remove any detrital and inorganic particles. Two subsamples of 500 μl supernatant were carefully removed from each Eppendorf tube and transferred to the flow cytometer's sample tubes. SybrGreen I solution (Molecular Probes, Eugene, OR, USA) was added for a final concentration of 2–4% to the flow cytometer's sample tubes before each measurement. The bacteria, flagellates and phytoplankton were determined according to the size classes measured with the flow cytometer. The size classes were

The solid hydrated density (ρs, g cm− 3) of the aggregates was determined in a density gradient using a modified version of previously reported methods (Feinberg and Dam, 1998; Schwinghamer, 1991). The density gradient consisted of seven dilutions which were made using Ludox TM colloidal silica, sucrose and distilled water. The dilutions had a density range between 1.05 and 1.43 g cm−3. The dilutions were buffered to pH 8.1 with 0.0125 M Tris plus 0.0125 M Tris–HCl (final concentration). Thus, the produced gradient was iso-osmotic with seawater of a salinity of 32. Two milliliters of each dilution was gently transferred to a 20 mL centrifuge tube with the densest dilution below and the least dense dilution on top. The density gradients were refrigerated overnight and allowed to adjust to the treatment temperature before use. One milliliter of seawater (salinity 32) was gently applied on top of each density gradient, and single aggregates were transferred to individual centrifugation tubes using a wide-tipped pipette letting the aggregates settle into the seawater layer without breaking. After a settling period of 2 to 4 h, the density gradients were centrifuged at 3000 rpm for 30 min to ensure that the aggregate had settled to the density layer equivalent to its solid hydrated density. One milliliter from the density layer containing the aggregate was removed from the tube, and its weight was measured on a Mettler Toledo fine balance. Additionally, 1 milliliter was sampled from all other density layers and investigated microscopically for aggregate constituents to ensure that the whole aggregate had settled to one density. Assuming iso-osmotic conditions between the aggregate and the density solution, the density of the removed gradient layer represents the solid hydrated density of the aggregate having neutral buoyancy within it. At each sampling time point, the solid hydrated densities of three aggregates from each treatment were measured except at the end time point of 380 h where the solid hydrated density of nine aggregates from each treatment were measured.

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27

21

2.9. Porosity

3.2. Aggregate dry weight

The porosity (p) of the aggregates was determined for each treatment on each sampling day according to the method given by Alldredge and Gotschalk (1988):

Dry weight (DW, μg) increased with increasing aggregate size in both treatments throughout the study and could be expressed as DW = 16.6 ESD1.98 (r2 = 0.76) and DW = 4.91 ESD2.1 (r2 = 0.68) for the ballasted and non-ballasted aggregates, respectively (data not shown). Power regressions were chosen to express DW as a function of aggregate size due to the fractal nature of the aggregates. There was a slight tendency for higher variation in the size-specific aggregate DW for the non-ballasted aggregates compared to the ballasted aggregates. This might be due to the incorporation of the small-sized clay particles into the ballasted aggregates, thereby resulting in more compact aggregates and more uniform structures compared to nonballasted aggregates. This is supported by the scanning electron microscope images of the ballasted and non-ballasted aggregates (Fig. 2); ballasted aggregates had high numbers of small smectite particles within them (Fig. 2B). The scanning electron microscope images of the freeliving samples (the ambient water samples) showed presence of smectite in the ballasted treatment at all the sampling time points (data not shown).

w ρs p ¼ 1− ; V

ð4Þ

where V is the volume of each aggregate in cm−3. Because the porosity is typically very high in marine aggregates, it is relatively insensitive to the density and dry weight of the aggregates (Alldredge and Gotschalk, 1988; Ploug et al., 2008a). 3. Results 3.1. Aggregate formation Aggregates were formed within the first 13 h of incubation in both treatments. Generally in both treatments, the abundance of larger aggregate sizes (N 2 mm) increased over time concurrent with a decrease in the abundance of small-sized aggregates (b 1 mm, Fig. 1). This indicated that the large aggregates scavenged the small aggregates over time. However, after 187 h and until 380 h of incubation, the smectite ballasted treatment showed an increase in the abundance of small aggregates (b1 mm) concomitant with a decrease in the abundance of large aggregates (N 3 mm), indicating that a disaggregation of the large aggregates occurred and resulted in the formation of small aggregates (b1 mm, Fig. 1B). The non-ballasted aggregates did not show any indication of disaggregation during the incubation period (Fig. 1A). Note that the non-ballasted aggregate abundance only ranges between 0 and 60 aggregates L−1 while the ballasted abundance ranges between 0 and 600 aggregates L−1.

3.3. Aggregate excess density, solid hydrated density, and porosity Due to the fractal nature of the aggregates, their excess density decreased with increasing size; i.e., porosity increased with increasing aggregate size. The decrease in excess density with increasing aggregate size was most pronounced for the ballasted compared to the nonballasted aggregates (Fig. 3A, B), which might be due to the lack of measurements of small aggregates (b 1 mm) in the non-ballasted treatment. We did not observe any change in size-specific excess density over time in the non-ballasted treatment (Fig. 3A) while we observed an increase in size-specific excess density for the ballasted aggregates on the last sampling day (Fig. 3B, 380 h). This might be caused by the disaggregation and re-aggregation of more compact aggregates after 187 h of incubation. We did not observe any significant differences in

Fig. 1. Aggregate evolution over time for the non-ballasted (A) and ballasted (B) treatments. The aggregate abundances were estimated for different size classes at the different time points during the whole incubation period. Note the different scales on the y-axes. The standard deviations are plotted as the positive error.

22

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27

aggregates measured on the last sampling day (380 h) showed higher size-specific sinking velocities in both treatments compared to the measurements done on the previous sampling days. The ballasted aggregates showed 2- to 3-fold higher size-specific sinking velocities on all the sampling days compared to the non-ballasted aggregates, and on average the ballasted aggregates were sinking 2-fold faster than the non-ballasted aggregates. The faster size-specific sinking velocities observed for the ballasted compared to the non-ballasted aggregates throughout the experimental period were due to the incorporation of smectite in the ballasted aggregates, which increased their compactness and density in comparison to the aggregates formed in the absence of smectite. The higher size-specific sinking velocities observed for the ballasted aggregates on the last sampling day (Fig. 3D) seem to have been caused by disaggregation and re-aggregation which occurred after 187 h (Fig. 1B), while the non-ballasted aggregates possibly became more dense over time due to degradation and possibly the hydrodynamic conditions within the roller tanks. Previous observations have also suggested that aging of marine snow aggregates can result in increased compactness and increased size-specific sinking velocities (Iversen and Ploug, 2010; Ploug et al., 2008a). 3.5. Particulate organic carbon and nitrogen content

Fig. 2. Scanning electron microscope images of particles within an aggregate from the non-ballasted (A) and ballasted (B) treatment. Smectite particles were observed attached to the spines of the diatoms within the ballasted aggregates (B). The non-ballasted aggregates did not have any smectite particles (A).

the solid hydrated density of the aggregated material within the aggregates between the two treatments (Table 1, Student's t-test; p = 0.991). The similar solid hydrated density between the two treatments shows that the higher size-specific excess density for the ballasted aggregates compared to the non-ballasted aggregates (Fig. 3A, B) was not due to a higher density of the smectite particles, but rather due to an increased compactness of the material within the ballasted aggregates. This is also supported by the higher size-specific DW for the ballasted aggregates (see Section 3.2) and the tendency for the ballasted aggregates to have lower porosities than the non-ballasted aggregates (Fig. 4). The differences in porosity between the two aggregate types were most pronounced on the first sampling day (Fig. 4; 48 h), however, during that day most of the ballasted aggregates were smaller than 1 mm while the non-ballasted aggregates were all larger than 1 mm (Fig. 1). Since porosities increased with increasing size, the large differences in porosities for the aggregates sampled on the first sampling day were more likely a result of the large size differences. 3.4. Aggregate sinking velocity Sinking velocity increased with increasing aggregate size for both the ballasted and the non-ballasted aggregates (Fig. 3C, D). Only the

Particulate organic carbon (POC) content in the aggregates increased with increasing aggregate size in both treatments (data not shown). We did not observe any differences in POC:PON ratios between the different sampling days in either of the treatments (One-Way Anova; p N 0.05) or between the two treatments on any of the sampling days (Student's t-test; p N 0.05) (data not shown). Therefore, the average POC:PON ratios for the entire experiment did not show any differences between the two treatments (Student's t-test; p = 0.22): the average POC:PON ratios were 9.5 ± 1.6 and 10.1 ± 0.7 for the non-ballasted and ballasted treatments, respectively. The average POC:DW ratios decreased in both treatments over the entire incubation time and decreased from 8 weight percent POC to 4 wt.% POC in the ballasted treatment, and from 31 wt.% POC to 8–9 wt.% POC in the non-ballasted treatment (Table 2). We observed significant differences between the POC:DW ratios between the ballasted and non-ballasted aggregates on all sampling days (Student's t-test; p b 0.05). 3.6. Microbial respiration rate within the aggregates The respiration rate per aggregate increased with increasing aggregate size in both treatments. The carbon respiration rate (Resp, μgC agg− 1 day− 1) could be expressed as Resp = 0.19 ESD1.2 (r2 = 0.36) and Resp = 0.10 ESD 1.3 (r2 = 0.32) for the nonballasted and ballasted aggregates, respectively (data not shown). The respiration rate increased proportionally to the POC content of the aggregates in both treatments, indicating first-order kinetics of POC degradation for both ballasted and non-ballasted aggregates (data not shown). The carbon-specific respiration rate was calculated by dividing the carbon respiration rate by the total POC content of each aggregate and was size-independent in both treatments at all times (Fig. 3E, F). There were no significant differences between the carbon-specific respiration rates for the two treatments on any of the sampling days or within any of the two treatments between the different sampling days (Student's t-test; p N 0.05). The overall average carbon-specific respiration rates for the two treatments throughout the whole experiments were not significantly different (Table 1, Student's t-test; p = 0.276). The average abundance of aggregate-associated and free-living bacteria, flagellates and phytoplankton did not indicate any influence from the presence of smectite (Fig. 5). For instance, the phytoplankton abundance for the ballasted treatment after 156 h of incubation was significantly larger than that for the non-ballasted treatment (Student's t-test: p = 0.03, Fig. 5D). However, at 187 h the average phytoplankton

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27

23

Fig. 3. Aggregate excess density (A, B), sinking velocity (C, D), carbon-specific respiration rate (C-spec resp rate) (E, F), and remineralization length scale L (G, H) are plotted against aggregate equivalent spherical diameter (ESD) for the non-ballasted (A, C, E, F) and ballasted treatments (B, D, F, H). The symbols for the time (in hours) of the different measurements are given in the legends in A. The legend is only valid for plots A–D. The carbon-specific respiration rate and remineralization length scale did not change throughout the experiment and the measurements for all sampling times were pooled and plotted together as solid black circles.

abundance in the ballasted treatment was lower than that observed at 156 h in the non-ballasted treatment, and at 350 h no significant difference was observed in average phytoplankton abundance between the Table 1 Average measurements (±SD) of solid hydrated density (ρs) of the composite particles within the aggregates, and carbon-specific respiration for the ballasted and non-ballasted aggregates over the whole incubation period. Treatment

Average ρs (g cm−3)

Average C-spec. resp. (d−1)

Ballasted Non-ballasted

1.26 ± 0.09 1.26 ± 0.10

0.12 ± 0.07 0.14 ± 0.06

two treatments (Student's t-test: p = 0.5, Fig. 5D). The aggregateassociated bacteria only showed significant differences between the two treatments after 350 h of incubation (Student's t-test: p = 0.01) with higher abundance in the non-ballasted treatment (Fig. 5B). The aggregate-associated flagellates only showed significant differences in abundance between the two treatments after 156 h of incubation with higher abundance in the non-ballasted treatment (Student's t-test: p b 0.01, Fig. 5F). No differences were observed for the free-living bacteria between the two treatments (Fig. 5A, C, E), except after 350 h of incubation for the free-living bacteria (Student's t-test: p b 0.01, Fig. 5A). However, the overall trend for both the free-living and the

24

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27

and sinking velocity (flux = concentration × sinking velocity), the 2-fold faster average sinking velocities of the ballasted aggregates compared to the non-ballasted aggregates suggests that the total POC flux of the ballasted aggregates was 2.7 times larger than that for the nonballasted aggregates (Table 3). 4. Discussion

Fig. 4. Average porosity for aggregates in the non-ballasted (black bars) and ballasted (white bars) treatments plotted for each measured time point (hours). Standard deviations are presented as positive error bars. The star marks significant differences between treatments at the level of p b 0.05 (Student's t-test).

aggregate-associated bacteria was that there were no clear differences between the two treatments (Fig. 5). 3.7. Remineralization length scale of aggregates The remineralization length scale, L (m− 1), was calculated by dividing the carbon-specific respiration rate by the settling velocity of the aggregates. This expresses the fractional remineralization of the organic matter within an aggregate per meter it settles. L decreased with increasing aggregate size for both treatments on all sampling days (Fig. 3G, H). The faster settling velocities observed for the ballasted aggregates compared to the non-ballasted aggregates resulted in lower L for the ballasted aggregates. L can be used as an index of the microbial carbon degradation or preservation within sinking organic aggregates. L was 3-fold lower within the ballasted aggregates compared to the nonballasted aggregates. 3.8. Ratios of total abundance, volume, mass and POC flux between the treatments To estimate if the ballasting by smectite could result in enhanced POC flux, we used the semi-quantitative estimates of aggregate abundance within the different size-classes (Fig. 1). These estimates provided good relative differences between the two treatments and offered a way to estimate the influence from ballasting on total POC flux. We observed that the average particle abundance in the ballasted treatment was 11-fold higher than in the non-ballasted treatment. However, since the ballasted aggregates were smaller than the nonballasted aggregates, there was no difference in the total volume of aggregated material between the two treatments (Table 3). The higher densities of the ballasted aggregates resulted in 8 times greater total aggregated DW of the ballasted compared to the non-ballasted aggregates. However, the total amount of aggregated POC was only 1.4 times larger in the ballasted compared to the non-ballasted treatment (Table 3). Since POC flux is the product of POC concentration Table 2 POC to dry weight (DW) ratios (± SD) over time for the ballasted and non-ballasted treatments. The POC:DW was calculated by dividing the aggregated POC content by the aggregate DW. Time since incubation start (h)

Ballasted POC:DW

Non-ballasted POC:DW

48 91 157 187 380

0.08 ± 0.02 0.06 ± 0.01 0.05 ± 0.01 0.04 ± 0.02 0.04 ± 0.003

0.31 ± 0.16 0.11 ± 0.05 0.11 ± 0.04 0.08 ± 0.01 0.09 ± 0.02

Observations from deep ocean sediment traps have shown strong correlations between fluxes of ballast minerals and organic carbon. This has led to the hypothesis that organic carbon is preserved within mineral ballasted aggregates due to increased aggregate density and sinking velocity and/or via protection of the organic matter due to association between the ballast minerals and the organic matter (Armstrong et al., 2002; Francois et al., 2002; Klaas and Archer, 2002). This and previous studies have only found support for the increased aggregate density and sinking velocity but not for any protection against degradation by the presence of ballast mineral (Iversen and Ploug, 2010; Ploug et al., 2008b). The smectite-ballasted aggregates differed from the non-ballasted aggregates by being more compact, having higher aggregate abundances, smaller maximum sizes, higher size-specific excess densities, lower porosities, and faster size-specific sinking velocities. These observations were not surprising and seem to be a general effect of incorporation of ballast minerals into settling organic aggregates (De La Rocha et al., 2008; Engel et al., 2009b; Hamm, 2002; Iversen and Ploug, 2010; Passow and De La Rocha, 2006; Passow et al., 2014; Ploug et al., 2008b). All direct measurements of size-specific sinking velocities of ballasted aggregates have shown an increase with size and with ballasting (Engel et al., 2009b; Iversen and Ploug, 2010; Passow et al., 2014; Ploug et al., 2008a,b; Schmidt et al., 2014), supporting the hypothesis that incorporation of ballast minerals into aggregates increases their density and sinking velocity. However, while Passow et al. (2014) observed higher excess densities for illite-ballasted aggregates, and hence higher size-specific sinking velocities, the decrease in average size of the ballasted aggregates resulted in lower average sinking velocities compared to those measured for non-ballasted aggregates. This shows that the outcome from mineral ballasting upon POC export depends on its effect on both the increase in abundance and excess density of the ballasted aggregates; i) will the increase in abundance make up for the mass-loss due to decreased maximum aggregate size, and ii) will the increase in excess density cause the smaller aggregates to sink as fast or faster than the larger non-ballasted aggregates? Our study shows that although the ballasted aggregates had 10 times higher abundances than the non-ballasted aggregates, their smaller sizes resulted in a similar total volume of aggregates and total aggregated amount of POC between the two treatments. Therefore, the 2-fold higher average sinking velocities for the ballasted aggregates suggested potential POC fluxes that were on average 2.7 ± 1.6 times larger for the ballasted aggregates compared to the non-ballasted, which is similar to a dust-induced increase in POC flux observed during a mesocosm study by Bressac et al. (2014). The 2- to 3-fold increase in total POC flux from ballasted compared to non-ballasted aggregates is similar to the suggested 3-fold better preservation of organic matter within ballasted aggregates, as indicated from the remineralization length scale (Fig. 3G, H). The similar average carbon-specific degradation rates in the two treatments indicated that the high potential POC flux for the ballasted aggregates was due to an increase in their size-specific sinking velocities compared to the non-ballasted aggregates. This supports previous suggestions that ballasting may enhance the vertical export of organic carbon by increasing the density and sinking velocity of the organic aggregates (Armstrong et al., 2002; Francois et al., 2002; Klaas and Archer, 2002). Still, our measurements were done in a flow chamber with laminar flow and without density discontinuities in the water. Aggregates settling through the water column will often encounter

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27

25

Fig. 5. Flow cytometer measurements of abundance of free-living (A, C, and E) and aggregate-attached (B, D, and F) bacteria (A–B), phytoplankton (C–D), and flagellates (E–F) measured at different timepoints throughout the experiment. The closed circles are for the non-ballasted treatments and the open circles are for the smectite ballasted treatments. Standard deviations are presented as positive and negative error bars. The abundance of free-living microorganisms was measured from ambient water samples from the roller tanks after allowing the aggregates to settle to the bottom of the tanks for a minimum of 30 min.

strong density discontinuity at the pycnocline, which may decrease their settling velocities (Alldredge et al., 2002; Kindler et al., 2010; Macintyre et al., 1995). This suggests that our sinking velocity measurements might overestimate those occurring in the upper few hundred meters of the water column where most density gradients are found. However, since small and dense aggregates have shorter retention times at density gradients than large and light aggregates (Kindler et al., 2010; Prairie et al., 2013), the influence from ballast minerals on export flux might be even greater in situ than that observed in our laboratory study. We observed no differences in average carbon-specific respiration rates over time or between the two treatments. Furthermore, the average carbon-specific respiration rates were similar to previous measurements within marine aggregates of different types

and sizes (Iversen et al., 2010; Iversen and Ploug, 2010; Ploug and Grossart, 2000; Ploug et al., 1999, 2008a,b), supporting the suggestion that carbon-specific respiration rates are similar within different marine particles irrespective of composition, size and type (Iversen and Ploug, 2010). The previous rates were measured within relatively fresh particles and, thus, only applied to the upper ocean while our study lasted 380 h (~ 16 days); during that time an aggregate settling with 200 m day− 1 would potentially reach a depth of ~ 3200 m. Since we observed high and similar carbon-specific respiration rates for both the ballasted and non-ballasted aggregates throughout the whole study period, the presence of ballast minerals did not have any influence on the microbial degradation of organic matter within the aggregates. This contradicts the suggestions that quantitative associations between

Table 3 Ratios of overall average measurements between the ballasted (BALL) and non-ballasted (NONBALL) treatments. The ratios were calculated with the different average values for the whole experiment period. The table shows the ratios in average total abundance of aggregates, total volume of aggregates, total dry weight of aggregates, total aggregated POC, and total POC flux. The total POC flux is calculated as POC flux = POC concentration × sinking velocity. Standard deviations are presented together with the ratios between the overall average values. Treatment

Total agg. abundances

Total agg. volume

Total agg. DW

Total agg. POC

Total agg. POC flux

Ratio of BALL:NONBALL

10.68 ± 6.97

1.25 ± 1.11

8.05 ± 6.77

1.38 ± 0.58

2.69 ± 1.58

26

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27

ballast minerals and organic matter protects the organic matter against degradation (e.g. Armstrong et al., 2002) and agrees with previous suggestions that ballasting of aggregates has a large influence on sinking velocities, but offers no protective mechanism against remineralization (Iversen and Ploug, 2010; Ploug et al., 2008a). We observed a decreasing trend in the POC:DW ratio within the ballasted aggregates, which seemed to stabilize at 4% after the first eight days of incubation (187 h). The initial decrease in the POC:DW may be explained by high degradation rates, which removed the organic matter within the aggregates but had no effect on the inorganic fraction. The stabilization of the POC:DW during the second week of incubations may be explained by the saturation capacity of organic aggregates for mineral particles, which has been found to be between 96 and 98% (De La Rocha et al., 2008; Passow and De La Rocha, 2006) corresponding to a 4 wt.% of organic matter. Observing the aggregate evolution for the ballasted aggregates (Fig. 1B), it seems that disaggregation occurred after 187 h and throughout the remaining incubation period. Hence, the strong correlation between POC and ballast minerals in the deep ocean (Armstrong et al., 2002; Francois et al., 2002; Klaas and Archer, 2002) might be expressed as a 3–7 wt.% of organic matter because disaggregation occurs when the organic matter content within the aggregates becomes less than 3–7% and can no longer “glue” the aggregates together. This would support the previous suggestion that POC to mineral ratios in the deep ocean are a function of the maximum amount of ballast mineral material that organic aggregates are able to scavenge and incorporate as they sink (Passow, 2004). The total amount of ballast minerals may be controlled by biotic processes such as degradation of organic matter or due to abiotic processes such as continuous scavenging of suspended minerals while the aggregates sink through the water column. Our study showed a direct increase in potential export flux from mineral ballasting of sinking aggregates. This was due to the sizespecific increased density and sinking velocity of ballasted aggregates while no changes were observed in POC content or microbial degradation of ballasted compared to the non-ballasted aggregates. Additionally, the smaller maximum aggregate size of the ballasted aggregates was compensated by an increase in total aggregate abundance whereby the total aggregated volume was similar between ballasted and nonballasted aggregates. This shows that clay particles can be an important ballast mineral that can increase aggregate sinking velocities and, potentially, export fluxes (Iversen and Ploug, 2010; Ploug et al., 2008a,b). A recent study found that the presence of the clay illite could not increase the amount of organic matter incorporated into sinking aggregates and suggested a less efficient biological pump in a dustier world (Passow et al., 2014). Our results offer the opposite outcome from the incorporation of clay into sinking aggregates as do several other ballast studies covering a range of ballast minerals (Bressac et al., 2014; Iversen et al., 2010; Iversen and Ploug, 2010; Ploug et al., 2008a,b; Ternon et al., 2010). If the ballasting of sinking aggregates does increase POC export, then the projected increased desertification and droughts in the future (e.g. Prospero and Nees, 1976) may lead to higher oceanic dust deposition and, thus, a more efficient biological pump. However, the interactions between ballast minerals and organic matter might not be that straightforward. For example, Ternon et al. (2010) observed that the deposition of Saharan dust in the Mediterranean Sea only caused an increase in the organic matter export when there was a simultaneous presence of organic matter and lithogenic material. It therefore seems that the results from the present study are mainly applicable to periods when dust deposition occurs directly after a phytoplankton bloom. 5. Concluding remarks This study shows that there is no protective mechanism against degradation of organic matter from the inclusion of the ballast mineral smectite for either freshly produced or aged marine snow aggregates. The main effect from ballasting with smectite was an increase in the

density of the aggregates and, thereby, an increase in the size-specific sinking velocities. Since the ballasted aggregates contained similar size-specific carbon content as non-ballasted aggregates, ballasting reduced the retention time of the settling aggregates in the surface ocean where high biological activity typically exists (Iversen et al., 2010; Jackson and Checkley, 2011; Stemmann et al., 2004). Thus, ballasting with smectite potentially governs high export of aggregates from the surface ocean into the deep. The deep ocean POC flux is characterized by quasi constant flux. Therefore, a large amount of the sinking organic matter making the descent through the upper 1000 m of the ocean is likely to reach the seafloor where the organic matter can be sequestered (Iversen et al., 2010; Iversen and Ploug, 2013). This implies that the incorporation of ballast minerals into sinking organic aggregates can increase the efficiency of the biological pump through their tendency to increase the size-specific sinking velocities of marine aggregates (see Iversen and Ploug, 2010; Ploug et al., 2008a,b). The findings from this study and previous studies might have important relevance for both the past and the future ocean. Recent studies of marine sediments collected from the Ocean Drilling Project have suggested that great ice-age dust loads may have increased carbon sequestration in the Southern Ocean (Martínez-Garcia et al., 2011). The authors suggested that the increased carbon sequestration was caused by an increased supply of iron and other essential limiting nutrients to the ocean from dust, which increased the productivity of the oceans and, thus, the carbon sequestration. However, this study suggests that the ballasting effect of the dust deposition could also have played an important role for the carbon sequestration by increasing the density of the settling aggregates and, thus, increasing the export production. This has important consequences for the future climate, in which it has been suggested that increased desertification and droughts will occur, which could lead to higher dust availability in the atmosphere and ocean (e.g. Prospero and Nees, 1976). The ballasting effect of ballast minerals could increase the efficiency of the biological pump in a dustier world.

Acknowledgements We thank Christiane Lorenzen for assistance during POC measurements. The oxygen microelectrodes were constructed by Gaby Eickert, Ines Schröder, and Karin Hohmann (Max Planck Institute for Marine Microbiology, Bremen). We thank Ulrike Jäekel, Tina Brenneis, and Kerstin Ötjen for help during the experimental period. We thank Uta Passow for discussions during the initial phase of the experiment planning. We further thank two reviewers and editors for valuable comments and suggestions to the manuscript. This study was supported by the Helmholtz Association (to MHI), the Alfred Wegener Institute for Polar and Marine Research (to MHI and MLR), the DFG-Research Center/Cluster of Excellence “The Ocean in the Earth System” (to MHI) and the Marie Curie Early Stage Training in Marine Microbiology (MarMic EST contract MEST-CT-2004-007776 to MLR). This publication is supported by the HGF Young Investigator Group SeaPump “Seasonal and regional food web interactions with the biological pump”: VH-NG-1000.

References Alldredge, A., Gotschalk, C., 1988. In situ settling behavior of marine snow. Limnol. Oceanogr. 33, 339–351. Alldredge, A.L., Granata, T.C., Gotschalk, C.C., Dickey, T.D., 1990. The physical strength of marine snow and its implications for particle disaggregation in the ocean. Limnol. Oceanogr. 35, 1415–1428. Alldredge, A.L., Gotschalk, C., Passow, U., Riebesell, U., 1995. Mass aggregation of diatom blooms: insights from a mesocosm study. Deep-Sea Res. II 42, 9–27. Alldredge, A.L., et al., 2002. Occurrence and mechanisms of formation of a dramatic thin layer of marine snow in a shallow Pacific fjord. Mar. Ecol. Prog. Ser. 233, 1–12. Armstrong, R.A., Lee, C., Hedges, J.I., Honjo, S., Wakeham, S.G., 2002. A new, mechanistic model for organic carbon fluxes in the ocean based on the quantitative association of POC with ballast minerals. Deep-Sea Res. II 49, 219–236.

M.H. Iversen, M.L. Robert / Marine Chemistry 175 (2015) 18–27 Arnarson, T.S., Keil, R.G., 2000. Influence of organic-mineral aggregates on microbial degradation of the dinoflagellate Scrippsiella trochoidea. Geochim. Cosmochim. Acta 69, 2111–2117. Avila, A., Queralt, I., Gallart, F., Martin-Vide, J., 1996. African dust over northeastern Spain: mineralogy and source regions. In: Guerzoni, S., Chester, R. (Eds.), The Impact of Desert Dust Across the Mediterranean. Kuwer Academic Publisher, pp. 201–206. Bressac, M., Guieu, C., Doxaran, D., Bourrin, F., Obloensky, G., Grisoni, J.-M., 2011. A mesocosm experiment coupled with optical measurements to assess the fate and sinking of atmospheric particles in clear oligotrophic waters. Geo-Mar. Lett. 32 (2), 153–164. Bressac, M., et al., 2014. Quantification of the lithogenic carbon pump following a simulated dust-deposition even in large mesocosms. Biogeosciences 11, 1007–1020. Broecker, W.S., Peng, T.H., 1974. Gas exchange rates between air and sea. Tellus 26, 21–35. Cole, T.G., Shaw, H.F., 1983. The nature and origin of authigenic smectites in some recent marine sediments. Clay Miner. 18, 239–252. De La Rocha, C.L., Passow, U., 2007. Factors influencing the sinking of POC and the efficiency of the biological carbon pump. Deep-Sea Res. II 54, 639–658. De La Rocha, C.L., Nowald, N., Passow, U., 2008. Interactions between diatom aggregates, minerals, particulate organic carbon, and dissolved organic matter: further implications for the ballast hypothesis. Glob. Biogeochem. Cycles 22, GB4005. http://dx.doi. org/10.1029/2007GB003156. Deuser, W.G., Ross, E.H., Anderson, R.F., 1981. Seasonality in the supply of sediment to the deep Sargasso Sea and implications for the rapid transfer of matter to the deep ocean. Deep-Sea Res. I 28, 495–505. Dilling, L., Alldredge, A.L., 2000. Fragmentation of marine snow by swimming macrozooplankton: a new process impacting carbon cycling in the sea. Deep-Sea Res. I 47, 1227–1245. Engel, A., et al., 2009a. Investigating the effect of ballasting by CaCO3 in Emiliania huxleyi, II: decomposition of particulate organic matter. Deep-Sea Res. II 56, 1408–1419. Engel, A., Szlosek, J., Abramson, L., Liu, Z., Lee, C., 2009b. Investigating the effect of ballasting by CaCO3 in Emiliania huxleyi: I. Formation, settling velocities and physical properties of aggregates. Deep-Sea Res. II 56, 1396–1407. Falkowski, P.G., Barber, R.T., Smetacek, V., 1998. Biogeochemical controls and feedbacks on ocean primary productivity. Science 281, 200–206. Feinberg, L.R., Dam, H.G., 1998. Effects of diet on dimensions, density and sinking rates of fecal pellets of the copepod Acartia tonsa. Mar. Ecol. Prog. Ser. 175, 87–96. Fischer, G., Karakas, G., 2009. Sinking rates and ballast composition of particles in the Atlantic Ocean: implications for the organic carbon fluxes to the deep ocean. Biogeosciences 6, 85–102. Fowler, S.W., Knauer, G.A., 1986. Role of large particles in the transport of elements and organic compounds through the oceanic water column. Prog. Oceanogr. 16, 147–194. Francois, R., Honjo, S., Krishfield, R., Manganini, S., 2002. Factors controlling the flux of organic carbon to the bathypelagic zone of the ocean. Glob. Biogeochem. Cycles 16 (4), 1087. http://dx.doi.org/10.1029/2001GB001722. Grossart, H.P., Ploug, H., 2001. Microbial degradation of organic carbon and nitrogen on diatom aggregates. Limnol. Oceanogr. 46, 267–277. Guieu, C., Ridame, C., Pulido-Villena, E., Bressac, M., Desboeufs, K., Dulac, F., 2014. Impact of dust deposition on carbon budget: a tentative assessment from a mesocosm approach. Biogeosciences 11, 5621–5635. Hamm, C.E., 2002. Interactive aggregation and sedimentation of diatoms and clay-sized lithogenic material. Limnol. Oceanogr. 47, 1790–1795. Hedges, J.I., Baldock, J.A., Gelinas, Y., Lee, C., Peterson, M., Wakeham, S.G., 2001. Evidence for non-selective preservation of organic matter in sinking marine particles. Nature 409, 801–804. Ingalls, A.E., Liu, Z., Lee, C., 2006. Seasonal trends in the pigment and amino acid compositions of sinking particles in biogenic CaCO3 and SiO2 dominated regions of the Pacific sector of the Southern Ocean along 170 degree W. Deep-Sea Res. I 53, 836–859. Iversen, M.H., Ploug, H., 2010. Ballast minerals and the sinking carbon flux in the ocean: carbon-specific respiration rates and sinking velocities of marine snow aggregates. Biogeosciences 7, 2613–2624 (doi:2610.5194/bg-2617-2613-2010). Iversen, M.H., Ploug, H., 2013. Temperature effects on carbon-specific respiration rate and sinking velocity of diatom aggregates — potential implications for deep ocean export processes. Biogeosciences 10, 4073–4085. Iversen, M.H., Poulsen, L.K., 2007. Coprorhexy, coprophagy, and coprochaly in the copepods Calanus helgolandicus, Pseudocalanus elongatus, and Oithona similis. Mar. Ecol. Prog. Ser. 350, 79–89. Iversen, M.H., Nowald, N., Ploug, H., Jackson, G.A., Fischer, G., 2010. High resolution profiles of vertical particulate organic matter export off Cape Blanc, Mauritania: degradation processes and ballasting effects. Deep-Sea Res. I 57, 771–784. Jackson, G.A., 1993. Flux feeding as a mechanism for zooplankton grazing and its implications for vertical particulate flux. Limnol. Oceanogr. 38, 1328–1331. Jackson, G.A., Checkley Jr., D.M., 2011. Particle size distribution in the upper 100 m water column and their implications for animal feeding in the plankton. Deep-Sea Res. I 58, 283–297. Kindler, K., Khalili, A., Stocker, R., 2010. Diffusion-limited retention of porous particles at density interfaces. PNAS 107, 22163–22168. Klaas, C., Archer, D.E., 2002. Association of sinking organic matter with various types of mineral ballast in the deep sea: implications for the rain ratio. Glob. Biogeochem. Cycles 16 (4), 1116. http://dx.doi.org/10.1029/2001GB001765. Koski, M., Kiørboe, T., Takahashi, K., 2005. Benthic life in the pelagic: aggregate encounter and degradation rates by pelagic harpacticoid copepods. Limnol. Oceanogr. 50, 1254–1263. Lampitt, R.S., Noji, T., Bodungen, B.V., 1990. What happens to zooplankton fecal pellets? Implications for material flux. Mar. Biol. 104, 15–23. Le Moigne, F.a.C., Gallinari, M., Laurenceau, E., De La Rocha, C., 2013. Enhanced rates of particulate organic matter remineralization by microzooplantkon are diminished by added ballast minerals. Biogeosciences 10, 5755–5765.

27

Lee, C., et al., 2009a. Particulate organic matter and ballast fluxes measured using time-series and settling velocity sediment traps in the northwestern Mediterranean Sea. Deep-Sea Res. II 56, 1420–1436. Lee, C., et al., 2009b. MedFlux: investigations of particle flux in the Twilight Zone. DeepSea Res. II 56, 1363–1368. Löder, M.G.J., Kraberg, A.C., Aberle, N., Peters, S., Wiltshire, K.H., 2011. Dinoflagellates and ciliates at Helgoland Roads, North Sea. Helgol. Mar. Res. http://dx.doi.org/10.1007/ s10152-10010-10242-z. Lunau, M., Lemke, A., Walther, K., Martens-Habbena, W., Simon, M., 2005. An improved method for counting bacteria from sediments and turbid environments by epifluorescence microscopy. Environ. Microbiol. 7, 961–968. Maas, L.R.M., 1994. On the surface area of an ellipsoid and related integrals of elliptic integrals. J. Comput. Appl. Math. 51, 237–249. Macintyre, S., Alldredge, A.L., Gotschalk, C.C., 1995. Accumulation of marine snow at density discontinuities in the water column. Limnol. Oceanogr. 40, 449–468. Mahowald, N.M., et al., 2009. Atmospheric iron deposition: global distribution, variability, and human pertubations*. Ann. Rev. Mar. Sci. 1, 245–278. Martínez-Garcia, A., Rosell-Melé, A., Jaccard, S.L., Geibert, W., Sigman, D.M., Haug, G.H., 2011. Southern Ocean dust-climate coupling over the past four million years. Nature 476, 312–316. Passow, U., 2004. Switching perspectives; do mineral fluxes determine particulate organic carbon fluxes or vice versa? Geochem. Geophys. Geosyst. 5, 1–5. Passow, U., De La Rocha, C., 2006. Accumulation of mineral ballast on organic aggregates. Glob. Biogeochem. Cycles 20, 1–7. Passow, U., Engel, A., Ploug, H., 2003. The role of aggregation for the dissolution of diatom frustules. FEMS Microbiol. Ecol. 46, 247–255. Passow, U., De La Rocha, C., Fairfield, C., Scmidt, K., 2014. Aggregation as a function of pCO2 and mineral particles. Limnol. Oceanogr. 59, 532–547. Ploug, H., 2001. Small-scale oxygen fluxes and remineralization in sinking aggregates. Limnol. Oceanogr. 46, 1624–1631. Ploug, H., Grossart, H.P., 2000. Bacterial growth and grazing on diatom aggregates: respiratory carbon turnover as a function of aggregate size and sinking velocity. Limnol. Oceanogr. 45, 1467–1475. Ploug, H., Jørgensen, B.B., 1999. A net-jet flow system for mass transfer and microsensor studies of sinking aggregates. Mar. Ecol. Prog. Ser. 176, 279–290. Ploug, H., Kuehl, M., Buchholz-Cleven, B., Jørgensen, B.B., 1997. Anoxic aggregates — an ephemeral phenomenon in the pelagic environment? Aquat. Microb. Ecol. 13, 285–294. Ploug, H., Grossart, H.P., Azam, F., Jørgensen, B.B., 1999. Photosynthesis, respiration, and carbon turnover in sinking marine snow from surface waters of Southern California Bight: implications for the carbon cycle in the ocean. Mar. Ecol. Prog. Ser. 179, 1–11. Ploug, H., Iversen, M.H., Fischer, G., 2008a. Ballast, sinking velocity, and apparent diffusivity within marine snow and zooplankton fecal pellets: implications for substrate turnover by attached bacteria. Limnol. Oceanogr. 53, 1878–1886. Ploug, H., Iversen, M.H., Koski, M., Buitenhuis, E.T., 2008b. Production, oxygen respiration rates, and sinking velocity of copepod fecal pellets: direct measurements of ballasting by opal and calcite. Limnol. Oceanogr. 53, 469–476. Ploug, H., Terbrüggen, A., Kaufmann, A., Wolf-Gladrow, D., Passow, U., 2010. A novel method to measure particle sinking velocity in vitro, and its comparison to three other in vitro methods. Limnol. Oceanogr. Methods 8, 386–393. Prairie, J.C., et al., 2013. Delayed settling of marine snow at sharp density transitions driven by fluid entrainment and diffusion-limited retention. Mar. Ecol. Prog. Ser. 487, 185–200. Prospero, J.M., Nees, R.T., 1976. Dust concentration in the atmosphere of the Equatorial North Atlantic: possible relationship to the Sahelian drought. Science 196, 1196–1198. Revsbech, N.P., 1989. An oxygen microsensor with a guard cathode. Limnol. Oceanogr. 34, 474–478. Sanders, R., Morris, P.J., Stinchcombe, M.C., Charalampopoulou, A., Lucas, M.I., Thomalla, S.J., 2010. Does ballast effect occur in the surface ocean? Geophys. Res. Lett. 37, 1–5. http://dx.doi.org/10.1029/2010GL042574. Schmidt, K., De La Rocha, C., Gallinari, M., Cortese, G., 2014. Not all calcite ballast is created equal: differing effects of foraminiferan and coccolith calcite on the formation and sinking of aggregates. Biogeosciences 11, 135–145. Schwinghamer, P., 1991. Seperation and concentration of living dinoflagellate resting cysts from marine sediments via density-gradient centrifugation. Limnol. Oceanogr. 36, 588–592. Stemmann, L., Jackson, G.A., Gorsky, G., 2004. A vertical model of particle size distributions and fluxes in the midwater column that includes biological and physical processes; Part II, application to a three year survey in the NW Mediterranean Sea. Deep-Sea Res. I 51, 885–908. Stokes, G.G., 1851. On the effect of the internal friction of fluids on the motion of pendulums. Camb. Philos. Soc. Trans. 9, 8–106. Ternon, E., et al., 2010. The impact of Saharan dust on the particulate export in the water column of the North Western Mediterranean Sea. Biogeosciences 7, 809–826. Thomalla, S.J., Poulton, A.J., Sanders, R., Turnewitsch, R., Holligan, P.M., Lucas, M.I., 2008. Variable export fluxes and efficiencies for calcite, opal, and organic carbon in the Atlantic Ocean: a ballast effect in action? Glob. Biogeochem. Cycles 22, 1–10. http://dx. doi.org/10.1029/2007GB002982. Thunell, R., Benitez-Nelson, C., Varela, R., Astor, Y., Muller-Karger, F., 2007. Particulate organic carbon fluxes along upwelling-dominated continental margins: rates and mechanisms. Glob. Biogeochem. Cycles 21, 1–12. http://dx.doi.org/10.1029/ 2006GB002793. Weaver, C.E., 1988. Clays, Muds, and Shales. Elsevier. White, F.M., 1974. Viscous Fluid Flow. 2nd ed. McGraw-Hill ed. Inc., New York.